Nicholas Coates, University of Ottawa

Nicholas Coates, Isaac Burns

Fighting & Responding

Background: The COVID-19 pandemic revealed a critical weakness in global health: viruses can spread globally within weeks, while antibody development requires months to years. Monoclonal antibodies are powerful for the detection and treatment of viral infections; however, discovery and optimization remain slow, expensive, and screening-intensive, limiting rapid-response and broader clinical deployment. AlphaFold3 represents a major advance by enabling accurate prediction of protein structures and interactions. Applied to antibody–antigen systems, it could shift antibody discovery toward computational design, compress timelines, reduce animal use, and expand access to biologics. Respiratory syncytial virus (RSV) is an ideal proof-of-concept target. Its F protein is targeted by palivizumab, the first licensed monoclonal antibody therapy for RSV. We have engineered an scFv version of palivizumab and established a high-throughput ELISA platform, providing a robust framework to test in silico predictions across many variants. Hypothesis: We hypothesize that AlphaFold3 can accurately predict interactions between antibody CDRs and viral antigens, enabling rational design of variants with altered binding properties. By integrating structural predictions with experimental binding data, we aim to build a machine-learning pipeline that guides antibody design across targets. Objectives: We aim to generate approximately 100 variant models of palivizumab–RSV F protein CDR interactions using AlphaFold3. Utilising the in-silico models, theoretical binding characteristics will be calculated and compared against wet-lab data. Expected outcomes: This project will generate the first dataset linking AlphaFold3-based predictions to experimental outcomes in a clinically relevant antibody–antigen system. If successful, this could allow for the near-immediate response to novel pathogens.